Loading [MathJax]/extensions/MathMenu.js
Intelligent Anti-Jamming Communication with Continuous Action Decision for Ultra-Dense Network | IEEE Conference Publication | IEEE Xplore

Intelligent Anti-Jamming Communication with Continuous Action Decision for Ultra-Dense Network


Abstract:

This paper addresses the issue of anti-jamming communication in dynamic and unknown environment for ultradense network (UDN). A deep reinforcement learning based antijamm...Show More

Abstract:

This paper addresses the issue of anti-jamming communication in dynamic and unknown environment for ultradense network (UDN). A deep reinforcement learning based antijamming algorithm is proposed by exploiting the frequency-hopping technology to cope with the jamming attack without estimating the mode and parameters of the jamming. In contrast to existing learning based schemes, in which the anti-jamming action, e.g., frequency hopping and transmit power adjustment, is taken from a pre-defined discrete anti-jamming strategy space, the proposed anti-jamming algorithm takes anti-jamming action from the continuous action space based on deterministic policy gradient. We represent this anti-jamming algorithm in an actor-critic framework, for which a convolution neural network (CNN) is adopted as the actor for anti-jamming action selection, and a deep neural network (DNN) is used as the critic for value estimation. We have implemented the proposed algorithm based on Google TensorFlow. Simulation results show that the system performance can be significantly improved by the proposed algorithm.
Date of Conference: 20-24 May 2019
Date Added to IEEE Xplore: 15 July 2019
ISBN Information:

ISSN Information:

Conference Location: Shanghai, China

Contact IEEE to Subscribe

References

References is not available for this document.